The distribution of phosphatidylinositol 4,5-bisphosphate in acinar cells of rat pancreas revealed with the freeze-fracture replica labeling method

Abstract

Phosphatidylinositol 4,5-bisphosphate [PI(4,5)P(2)] is a phospholipid that has been implicated in multiple cellular activities. The distribution of PI(4,5)P(2) has been analyzed extensively using live imaging of the GFP-coupled phospholipase C-δ1 pleckstrin homology domain in cultured cell lines. However, technical difficulties have prevented the study of PI(4,5)P(2) in cells of in vivo tissues. We recently developed a method to analyze the nanoscale distribution of PI(4,5)P(2) in cultured cells by using the quick-freezing and freeze-fracture replica labeling method. In principle, this method can be applied to any cell because it does not require the expression of artificial probes. In the present study, we modified the method to study cells of in vivo tissues and applied it to pancreatic exocrine acinar cells of the rat. We found that PI(4,5)P(2) in the plasma membrane is distributed in an equivalent density in the apical and basolateral domains, but exists in a significantly higher concentration in the gap junction. The intracellular organelles did not show labeling for PI(4,5)P(2). The results are novel or different from the reported distribution patterns in cell lines and highlight the importance of studying cells differentiated in vivo.

Figure 1. Outline of the modified QF-FRL method for the analysis of cells of in vivo tissues.

A small piece of tissue was placed in an aluminum disc and covered with a thin stainless steel foil. The assembly was frozen using high-pressure freezing, and freeze-fracture replicas were prepared. The replicas were treated sequentially with trypsin, DNase I, and SDS and then labeled. PI(4,5)P₂ was marked using colloidal gold particles and observed with electron microscopy.

Figure 2. PI(4,5)P₂ in rat pancreatic acinar cells labeled by the QF-FRL method.

(A) PI(4,5)P₂ was labeled positively only in the P face of the plasma membrane [PM (P-face)], whereas the E-face [PM (E-face)] was devoid of labeling. The P and E faces correspond to the inner and outer leaflet, respectively. Intracellular organelles, including secretory granules (SG), were not significantly labeled. The inset shows a small portion of the PM P face, and some gold particles are indicated with arrows. (B) A control experiment. GST-PH (K30N, K32N), a mutant that does not bind PI(4,5)P₂, was used at the same concentration as GST-PH. GST-PH (K30N, K32N) produced little labeling. The inset is an enlargement of the PM P face to show the virtual absence of gold particles. (C) Quantification of the relative labeling density in the plasma membrane (average ± standard error). The density of colloidal gold particles is shown as the relative ratio to that of the P face labeled by GST-PH. The data were collected from three independent experiments, and areas greater than 7.5 µm² were analyzed. The results show that only the P face is labeled specifically with GST-PH. The labeling density in the E face was not statistically different between samples treated with GST-PH and GST-PH (K30N, K32N).

Figure 3. PI(4,5)P₂ in the plasma membrane.

(A) PI(4,5)P₂ was labeled in an equivalent density in the P face of the apical domain (Ap), the basolateral domain (BL), and the tight junctional region (TJ) of the plasma membrane. The tight junctional region is demarcated by dotted lines. In the apical membrane, most microvilli were fractured at the basal portion and seen as stubs (arrowheads), but some were fractured longitudinally (arrows; arrows in the inset). The labeling density in the microvillar membrane was not significantly different from that in the flat portion of the apical membrane. (B) A high magnification image of the tight junctional region (the area in the rectangle of Fig. 3A). The tight junction was observed as shallow grooves in the P face (arrows). Some PI(4,5)P₂ labeling was observed near the groove, but the total density in the junctional region was not different from that of the apical and basolateral membrane domains. (C) Quantification of the relative labeling density in the P face (average ± standard error). The density of colloidal gold particles is shown as the relative ratio to that of the apical domain. The data were collected from three independent experiments, and the total measured areas were 4.6 µm² (apical), 8.6 µm² (basolateral), and 7.7 µm² (tight junction). The labeling density was equivalent in the three regions. GST-PH (K30N, K32N) produced little labeling in all three domains.

Figure 4. PI(4,5)P₂ in the gap junction.

(A) PI(4,5)P₂ was labeled more intensely in the P face of the gap junction (GJ) than that of the surrounding undifferentiated basolateral membrane. The E face of the gap junction was not labeled. The gap junctional plaques are circumscribed by dotted lines. (B) Quantification of the relative labeling density in the P face (average ± standard error). The data were collected from pairs consisting of a gap junctional plaque and its surrounding membrane area, in four independent experiments. The labeling density in the gap junction is significantly higher than that of the undifferentiated membrane (Student's t-test, *p = 0.04). GST-PH (K30N, K32N) gave only negligible labeling both in the gap junction and in the surrounding membrane.

Figure 5. PI(4,5)P₂ in intracellular organelles.

(A-E) Organelles were identified by morphological criteria. The nuclear membrane (Nuc) (A), the ER (B), the Golgi apparatus (C), the secretory granule (SG) (A and D), and the mitochondrion (Mit) (E) were devoid of specific labeling for PI(4,5)P₂. More colloidal gold particles were seen in the P face of the secretory granule than the other organelles, but they were observed in a similar density even when GST-PH was replaced with GST-PH (K30N, K32N) (Fig. 2B). The inset (A) is an enlargement of a small portion of the P face of the secretory granule membrane to show the presence of non-specific labeling (arrows). (F) Quantification of the relative labeling density (average ± standard error). The labeling density in the secretory granule is shown as the relative ratio to that of the plasma membrane. The data were collected from three independent experiments and areas more than 9.3 µm² was measured. An equivalent number of colloidal gold particles were observed in the P face of the secretory granule when either GST-PH or GST-PH (K30N, K32N) was used. Therefore, we concluded that the labeling in the secretory granule was insignificant.